In the Search for Reliable Biomarkers for the Early Diagnosis of Autism Spectrum Disorder: the Role of Vitamin D

Metabolic Brain Disease (2018) 33:917–931 https://doi.org/10.1007/s11011-018-0199-1

ORIGINAL ARTICLE

In the search for reliable biomarkers for the early diagnosis of autism spectrum disorder: the role of vitamin D

Afaf El-Ansary1,2,3,4 & John J. Cannell5 & Geir Bjørklund6 & Ramesa Shafi Bhat7 & Abeer M. Al Dbass7 & Hanan A. Alfawaz8 & Salvatore Chirumbolo9 & Laila Al-Ayadhi3,4,10

Received: 13 May 2017 /Accepted: 2 February 2018 /Published online: 1 March 2018 # Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract Autism spectrum disorder (ASD) affects about 1% of the world’s population. Vitamin D is thought to be essential for normal brain development and modulation of the immune system. Worldwide about 1 billion people are affected by vitamin D deficiency. High-sensitivity C-reactive protein (hs-CRP), cytochrome P450 2E1 (CYP2E1) and 8- hydroxy-2′-deoxyguanosine (8-OH-dG) are biomarkers related to inflammation and oxidative stress. In the present study, these biomarkers were together with serum 25-hydroxyvitamin D (25(OH)D3) analyzed in 28 (mean age seven years) Saudi male patients with ASD. The study was conducted to determine if there is any relationship between vitamin D levels, the tested biomarkers and the presence and severity of ASD. The hope was to identify if these biomarkers may be useful for early ASD diagnosis. The Childhood Autism Rating Scale (CARS) and the Social Responsiveness Scale (SRS) were used to measure autism severity. The results of the ASD children were compared with 27 age and gender-matched neurotypical controls. The data indicated that Saudi patients with ASD have significantly lower plasma levels of 25(OH)D3 than neurotypical controls (38 ng/ml compared to 56 ng/ml, respectively; [P = 0.001]). Surprisingly, the levels of CYP2E1 were lower in the children with ASD than the neurotypical controls (0.48 ± 0.08 vs. 69 ± 0.07 ng/ml, respectively; P = 0.001). The ASD children also had significantly higher levels of hs-CRP (0.79 ± 0.09 vs. 0.59 ± 0.09 ng/ml, respectively; P = 0.001) and 8-OH-dG (8.17 ± 1.04 vs. 4.13 ± 1.01 ng/ml, respectively; P = 0.001, compared to neurotypical age and gender-matched controls. The values for hs- CRP and 8-OH-dG did not correlate [P < 0.001] with autism severity. There was found a relationship between autism severityontheCARSscaleandthelevelsof25(OH)D3 and CYP1B1. But this was not found for SRS. All four biomarkers seemed to have good sensitivity and specificity, but the sample size of the present study was too small to determine clinical usefulness. The findings also indicate that inadequate levels of vitamin D play a role in the etiology and severity of autism. Furthermore, the results of the present study suggest the possibility of using

25(OH)D3, CYP1B1, hs-CRP and 8-OH-dG, preferably in combination, as biomarkers for the early diagnosis of ASD.However,furtherresearchisneededtoevaluatethishypothesis.

* Geir Bjørklund 6 Council for Nutritional and Environmental Medicine, Toften 24, [email protected] 8610 Mo i Rana, Norway

7 Biochemistry Department, Science College, King Saud University, 1 Central Laboratory, Female Centre for Scientific and Medical Riyadh, Saudi Arabia Studies, King Saud University, Riyadh, Saudi Arabia 8 2 Department of Food Science and Human Nutrition, King Saud Medicinal Chemistry Department, National Research Centre, Dokki, University, Riyadh, Saudi Arabia Cairo, Egypt 3 Autism Research and Treatment Center, Riyadh, Saudi Arabia 9 Department of Neurological and Movement Sciences, University of Verona, Verona, Italy 4 Shaik AL-Amodi Autism Research Chair, King Saud University, Riyadh, Saudi Arabia 10 Department of Physiology, Faculty of Medicine, King Saud 5 Vitamin D Council, San Luis Obispo, CA, USA University, Riyadh, Saudi Arabia 918 Metab Brain Dis (2018) 33:917–931

Keywords Autism . Vitamin D . 8-OHdG . CYP1B1 . High-sensitivity C-reactive protein . Autism biomarkers

Introduction Chirumbolo et al. 2017). Vitamin D is produced endogenously in the skin when ultraviolet light triggers its synthesis. For Autism spectrum disorder (ASD) is characterized by impaired most people in the world, about 90% of the vitamin D is social interaction, problems with verbal and nonverbal com- produced in this way, and about 10% is coming from food munication, and repetitive behaviors. Currently at least one of and dietary supplements (Holick et al. 2011; Mahmoud and 68 children in the US has ASD (Zablotsky et al. 2015; Ali 2014). However, it is important for people living in high Christensen et al. 2016). In the Gulf States, the prevalence of latitudes to get enough vitamin D from the diet. In Norway, the ASD ranges from 1.4 to 29 per 10,000 persons, therefore still primary vitamin D sources in the diet are fatty fish, fortified representing a great health concern (Salhia et al. 2014). butter and margarine (Løken-Amsrud et al. 2012), and supple- Diagnosis of ASD may appear problematic, especially early mentation with cod liver oil (Saltytė Benth et al. 2012). in the child’s life. According to recent studies, the average age Vitamin D fortified milk and infant formula contain signifi- at diagnosis of ASD in the US is 53 months (ADDM Network cant amounts of vitamin D for infants, but it is not clear that 2014). Earlier diagnosis appears to improve the prognosis at even 400 IU/day (10 μ g) is considered as an adequate amount least in higher functioning children (Anderson et al. 2014). In (Gross et al. 2013). this perspective, clinically useful biomarkers are desperately Whatever the source, vitamin D is biologically inert in the needed, as they may improve ASD diagnosis at the earliest, human body and needs two hydroxylations to be activated. although few molecules can be suitable for the purpose, such The first of these is a liver hydroxylation that converts 25- as the biomarkers of the glutamate excitotoxicity and oxida- hydroxyvitamin D (25(OH)D3, calcidiol). This is the primary tion found in ASD (El-Ansary 2016). Several authors have circulating form of vitamin D. Next, the metabolite 25(OH)D3 proposed various biomarkers of increased oxidative stress is hydroxylated in the kidneys and by some other body cells to and impaired methylation capacity, such as methionine, ho- 1,25-dihydroxy vitamin D (1,25(OH)2D3, calcitriol), which is mocysteine, cystathionine, cysteine, total glutathione, S- a secosteroid hormone. A CYP enzyme (CYP27B1) catalyzes adenosyl-L-methionine, adenosine, and oxidized glutathione the hydroxylation of 25(OH)D3 to its active hormonal form (James et al. 2004; James 2013; Hodgson et al. 2014). El- (calcitriol), and another two CYP enzymes (CYP2R1 and

Ansary et al. (2016) used libraries of biomarkers to predict CYP24A1) metabolize vitamin D3 to 25(OH)D3 (Schuster sensory abnormities in ASD accurately. In one study, all the 2011). The half-life of 25(OH)D3 varies, but the half-life is 30 children with ASD undergoing the investigation, showed relatively long. Its half-life has been found to be 2–3weeks increased urinary kininogen-1 levels, when compared to when in both a high ultraviolet B atmosphere (Gambia) and a neurotypical controls (p < 0.001). Interleukin-6 and serotonin low one (the UK) (Jones et al. 2015; Holick 2006). have also been proposed as biomarkers with enough sensitiv- Plasma vitamin D metabolites are mostly transported by ity and specificity to be useful, but sample size and failure of vitamin D-binding protein (DBP), also known as group- replication limited their clinical usefulness (Yang et al. 2015). specific component (GC). Albumin is the secondary vitamin Several other studies that have looked at biomarkers in ASD D binding protein, especially in patients with a low DBP con- have also had too small sample sizes to determine specificity centration (Bikle et al. 1985;Currenti2010). In addition to and sensitivity accurately (El-Ansary et al. 2011a;Mingetal. vitamin D transport, DBP is an essential glycoprotein 2012; Emond et al. 2013;Kuwabaraetal.2013). Yusuf and macrophage-activating factor (GcMAF). It also scavenges ac- Elsabbagh (2015) warned against the use of biomarkers to tin and has fatty acid binding properties (Carpenter et al. diagnose ASD due to the lack of adequate sensitivity and 2013). Genotypes of DBP have been found to correlate with specificity in the current literature, so refreshing a concern still differences in plasma DBP and 25(OH)D3 levels (Chun et al. occurring in the research on autism pathology. ASD, 2016). An interaction between vitamin D status and some which is regarded as a multifactorial disorder, is yet maternal variants of DBP (GC) as well as birth weight heterogeneous in its presentation. Therefore, it is unlike- (Delanghe et al. 2015) has been reported. A high metabolic ly that a single biomarker would be helpful to diagnose activity of GcMAF has been demonstrated in the blood of the disorder. A panel of biomarkers for ASD would patients with ASD (Siniscalco et al. 2014). have to be found that have enough specificity and sen- Research indicates that vitamin D protects cells by up- sitivity to be clinically useful (Pelphrey 2017). regulating DNA-repair genes that stabilize the genome and Existing research indicates that vitamin D plays a crucial helps in the protection against oxidative stress and toxic ef- role in human health beyond the regulation of calcium homeo- fects (Gonzalez-Suarez et al. 2011;Grazianoetal.2016). The stasis and bone mineralization (Holick 2007; Bjørklund 2016; most common genetic finding reported in ASD is a wide- Metab Brain Dis (2018) 33:917–931 919 spreading of de novo mutations, the location of which varies Furthermore, one study showed that siblings with ASD had from child to child and does not allow a full elucidation of lower 25(OH)D3 levels at birth (Fernell et al. 2015). A autism genomics. Evidence showed that vitamin D prevents Swedish population study (N = 509,639) found an association and repairs these small mutations (Kinney et al. 2010). between diagnosed lifetime maternal vitamin D deficiency (< Interestingly, vitamin D is also involved in the glutathione 10 ng/ml) and risk of ASD, especially of ASD with intellec- cycle through its upregulation of γ-glutamyl transpeptidase, tual disability (Magnusson et al. 2016). In another population which is the rate-limiting enzyme in the glutathione metabolic study (n = 4229), Australian women with low vitamin D pathway (Garcion et al. 1996, 2002). Many children with levels (< 10 ng/ml) in mid-gestation had at birth twice as much ASD have low glutathione levels (Ghanizadeh et al. 2012). risk to bear a child with autistic traits compared to women The nuclear activity of vitamin D depends on the interaction with vitamin D levels >20 ng/ml (Vinkhuyzen et al. 2016). with vitamin D receptor (VDR). Chinese researchers investigated 68 children diagnosed with There are vitamin D receptors in most areas of the brain. ASD, and 68 sex and age-matched neurotypical children Vitamin D appears to protect neurons and promote neural (Chen et al. 2016). They found that women with the lowest growth (Ali et al. 2016). Vitamin D deficiency has been asso- quartile of vitamin D levels in their first trimester were ciated with cognitive pathology, including ASD (Cherniack four times more likely to give birth to a child with and Troen 2011;Mazaheryetal.2016). It has been ASD. In the same study, first-trimester vitamin D levels hypothesized that gestational and early childhood vita- were associated with autism severity (R = −0.302; P = min D deficiency might be a cause of ASD (Cannell 0.001) (Chen et al. 2016). 2008). Further, it has been suggested that high-dose vi- One of the most frequent findings in ASD is a widespread tamin D treatment has beneficial effects in ASD sub- de-novo genetic damage (Kinney et al. 2010). DNA damage is jects (Cannell 2013;Saadetal.2016a). a critical cause of cellular dysfunction and cell death. For Vitamin D deficiency during pregnancy is widespread and example, 8-hydroxydeoxyguanosine (8-OH-dG) is the prod- results in infants who have 33–50% less vitamin D than their uct of free radical attack to DNA-bound guanosine. It is the mothers (Wegienka et al. 2016). In a mostly white cohort in most abundant product of cellular DNA oxidation (Helbock Iowa, 70% of 4-month-old breastfed infants had mean et al. 1999) and is a potent mutagen (Calderón-Garcidueñas

25(OH)D3 < 12 ng/ml (ideal range 40–60 ng/ml) (Dawodu et al. 1999) and hence a marker for oxidative damage from and Wagner 2012). The prevalence of levels <15 ng/ml in DNA. As indicators that ASD might be an inflammatory dis- the infants was 50% in the summer and 79% in the winter order, reduced glutathione, increased malondialdehyde (a bio- periods of the year. Fifty-seven percent of the infants who marker for oxidative stress), and higher 8OHdG may be found were followed for six months still had vitamin D deficiency. in ASD children compared with neurotypical controls (Sajdel- In a study from Cincinnati, 18% of exclusively breastfed in- Sulkowska et al. 2008; Sajdel-Sulkowska et al. 2009). fants aged 1 month had vitamin D levels <10 ng/ml (in the Sajdel-Sulkowska et al. (2009)reportedanincreasein8- rachitic range); 76% of the infants and 17% of their mothers OH-dG in cerebella of ASD individuals, a finding that later had serum 25(OH)D3 < 20 ng/ml (Dawodu et al. 2014). The was confirmed by Rose et al. (2012). However, two more recommended range of vitamin D by the Endocrine Society is recent studies failed to find elevated 8-OH-dG in individuals 40–60 ng/ml (Holick et al. 2011). In Boston, 58.0% of new- with ASD (Ming et al. 2005; Yui et al. 2016). Chauhan et al. borns and 35.8% of the mothers had 25(OH)D3 <20ng/mL (2004) found remarkably increased urinary levels of 8-OH-dG and 38.0% of the infants and 23.1% of the mothers had in ASD patients. Also, children with brain damage have been

25(OH)D3 <15ng/mL. found to have increased urinary 8-OH-dG levels (Fukuda et al. If vitamin D deficiency were one of the causes of ASD, 2008). In a randomized controlled clinical trial, even the small then ASD would be common in children with rickets. This has dose of 800 IU/day of vitamin D was enough to reduce 8-OH- first relatively recently been studied. Zaky et al. (2015) dG by 25% (Fedirko et al. 2010). reported that 25% of children with rickets also have The cytochrome CYP-450 (CYP) family of enzymes is ASD. A recent meta-analysis of 11 studies found signif- responsible for anabolism and catabolism of both exogenous icantly lower 25(OH)D3 levels measured in ASD chil- and endogenous materials (Nebert and Russell 2002). Studies dren compared to neurotypical controls (Wang et al. have shown that increased CYP1B1 plays a significant role in 2016). Several studies have reported an inverse correla- oxidative stress (Dey 2013). Increased CYP1B1 is associated tion between the levels of vitamin D and autism severity with increased oxidative stress, mainly in alcoholic liver dis- measured by standard rating scales (Mostafa and Al- ease. People who are co-exposed to toluene and tobacco Ayadhi 2012;Gongetal.2014;Duetal.2015). smoke are due to a possible CYP1B1 repression at higher risk VitaminDdeficiencymayinsomechildrenwithASD of oxidative stress (Jiménez-Garza et al. 2015). Also, contribute to the production of serum anti-MAG auto- CYP1B1 gene mutation is associated with the abnormal antibodies (Mostafa and Al-Ayadhi 2012). social-emotional behavior as autistic phenotype (Chakrabarti 920 Metab Brain Dis (2018) 33:917–931 et al. 2009). Moreover, abnormal CYP1B1 is correlated with and the Ethics Committee at King Khalid Hospital, King Saud altered lipid metabolism repeatedly reported in ASD patients University approved the study. (Lin et al. 2015). High-sensitivity C-reactive protein (hs-CRP) is a highly Evaluation of autism severity conserved plasma protein that participates in the systemic response to some forms of inflammation. Some, but not all, In the present study, the Childhood Autism Rating Scale studies have found elevated hs-CRP in children with ASD (CARS) and the Social Responsiveness Scale (SRS) were as a marker of inflammation, which is thought to be a mech- used to measure autism severity. CARS rates the child on a anism involved in the neuropathology of ASD (Singh 2005). scale from one to four in each of 15 areas relating to people Mellenthin et al. (2014) reported that different inflammatory (emotional response; imitation; body use; object use; listening biomarkers are differently associated with 25(OH)D3.They response; fear or nervousness; verbal communication; non- alsofoundaninterestingU-shapedassociationbetween verbal communication; activity level; level and reliability of

25(OH)D3 and hs-CRP (Mellenthin et al. 2014). intellectual response; adaptation to change; visual response; It was hypothesized that serum 25(OH)D3 might be related taste, smell and touch response; and general impressions). to cytochrome P450 1B1 (CYP1B1), high-sensitivity C-reac- Children who score 30–36 on the CARS scale have mild to tive protein (hs-CRP), and 8-OH-dG in Saudi patients with moderate autism, while those with scores ranging between 37 ASD compared to neurotypical controls. If a relationship were and 60 points have severe autism (Mick 2005). On the SRS, a found, this could help to better understand the etiological total score of 76 or higher is considered very severe autism mechanism and the possible role of vitamin D in the control and strongly associated with a clinical diagnosis of autistic of oxidative stress and inflammation in ASD. In the present disorder. A score of 60–75 is interpreted as mild to moderate study, it was expected to find significantly elevated levels of range of social impairment (Constantino et al. 2003). hs-CRP, CYP1B1, and 8-OH-dG, as well as a decreased level of 25(OH)D3 in individuals with ASD, compared to Blood sampling and processing neurotypical controls. After an overnight fast, blood samples were collected in the morning from both groups. The samples were taken from the Materials and methods antecubital vein in 10 mL tubes containing sodium heparin. The tubes were at room temperature and centrifuged at Subjects 3500 rpm for 15 min within the first 4 h following the blood withdrawal. Plasma and red blood cells were obtained and The present cross-sectional study was performed on 28 male eventually stored at −80 °C until analysis. children with ASD. The patients were recruited from the Autism Research and Treatment Center at the Faculty of Biochemical analysis Medicine, King Saud University in Riyadh, Saudi Arabia.

Of these individuals, 18 were verbal, and 10 were nonverbal. The plasma concentration of vitamin D3, 25(OH)D3,was Their ages ranged between 3 and 12 years (mean SD = 7.0 ± measured by a commercially available HPLC kit (Eagle 2.34 years). The patients were diagnosed using the Diagnostic Biosciences, Nashua, NH, USA) according to the manufac- and Statistical Manual of Mental Disorders, 4th Edition, Text turer’s instructions. 400 μL sample, calibration or control, was Revision (DSM-IV-TR) criteria (APA 2000). The control mixed with another 400-μL ice-cold internal standard, and group comprised 27 age- and sex-matched neurotypical chil- briefly mixed. The sample was then mixed with 500 μLpre- dren with a mean age of 7.2 ± 2.14 years. The controls were cipitation reagent (PREC) and centrifuged at 10.000 rpm for normally developing, neurotypical children, unrelated to the 5 min. For HPLC injection 50 μl of the upper liquid layer was ASD subjects. The control children were neurotypical older used. The HPLC separation uses Breversed phase^ (IC3401rp) siblings of healthy infants who were attending the Well Baby column at 30 °C. Chromatograms are produced by UV-detec- Clinic at King Khalid University Hospital for a routine check- tor. The separation took 15 min. A plasma calibrator (Eagle up of their growth parameters. They had no clinical indica- Biosciences, Nashua, NH, USA) was used to quantify the tions of infectious disease or neuropsychiatric disorders. results. The internal standard method was used to calculate

Participants were excluded from the study if they had a diag- the concentration of 25(OH)D3 via integration of peak and nosis of fragile X syndrome, epileptic seizures, obsessive- height areas, respectively. The results were expressed as ng/ compulsive disorder, affective disorders, or any additional mL. Vitamin D deficiency was defined as serum 25(OH)D3 psychiatric or neurological diagnoses. All the participants levels <20 ng/mL and insufficiency at serum levels <30 ng/ had normal results for urine analysis and sedimentation rate. mL; knowing that the Endocrine Society recommends levels Written consent was obtained from the parents of each subject, between 40 and 60 ng/ml (Holick et al. 2011). Metab Brain Dis (2018) 33:917–931 921

Assay of cytochrome P450 1B1 Statistical analysis

Quantitative detection of cytochrome P450 1B1 (CYP1B1) in Data were analyzed using IBM SPSS 16 statistical software. plasma was done with an ELISA kit (Biomatik Corporation, The results were expressed as mean ± SD, and all statistical Cambridge, Ontario, Canada). The microtiter plate provided comparisons were made using an independent t-test with P ≤ in this kit is pre-coated with an antibody specific to CYP1B1. 0.05 considered as significant, through a Wilcoxon-Mann Standards or samples are added to the appropriate microtiter Whitney test when data were not normally distributed plate wells with a biotin-conjugated antibody preparation spe- (Shapiro-Wilk’s test negative). For normal distribution data cific for CYP1B1. Next, avidin-conjugated to horseradish per- were elaborated with an ANOVA (Tukey’s post hoc test). oxidase (HRP) is added to each microplate well and incubat- The utility diagrams of the measured parameters were drawn ed. After 3,3′,5,5′ tetramethylbenzidine (TMB) substrate solu- with the X-axis representing percentile rank of the biomarker, tion is added, only those wells that contained CYP1B1, biotin- the Y-axis representing the probability of identifying the dis- conjugated antibody, and enzyme-conjugated avidin exhibited ease, and the horizontal line the prevalence of the disease. a change in color. The enzyme-substrate reaction is terminated Multiple regression analysis was also performed using vita- by the addition of sulphuric acid solution, and the color min D as a dependent variable and CRP, CYP1B1 and 8-OH- change is measured spectrophotometrically via wavelength dG as independent variables. Qualitative variables were elab- of 450 nm ± 10 nm. The concentration of CYP1B1 in the orated using covariance analysis (Marusteri and Bacarea samples is then determined by comparing the optical density 2010). Predictiveness curves for the risk models were calcu- of the samples to the standard curve. Detection range is 0.156– lated on IBM SPSS 16 and based on existing literature (Pepe 10 ng/mL. et al. 2008; Huang et al. 2007; Vickers and Elkin 2006).

Assay of high-sensitivity C-reactive protein Results High-sensitivity C-reactive protein (hs-CRP) was measured using a solid phase direct sandwich diagnostic kit (GenWay Table 1 shows that when autism was described as mild or Biotech, San Diego, CA), according to the manufacturer’s severe in CARS but not in SRS, the suggested markers, i.e., instruction. The samples and anti-CRP-HRP conjugate were vitamin 25(OH)D3, hs-CRP, CYP1B1 and 8-OH-dG indicated added to the wells coated with monoclonal antibody (mAb) to a significant marked difference with respect to neurotypical CRP. Hs-CRP in the patient’s plasma sample binds to anti- children (control) (Table 1). In severe autistic subjects, as

CRP mAb on the well and the anti-hs-CRP second antibody emerging from CARS, 25(OH)D3 plasma levels were then binds to hs-CRP. Unbound protein and HRP conjugate 57.45% lower than controls, hs-CRP higher of about 34%, are washed off with wash buffer. Upon the addition of the CYP1B1 decreased as expected (about 70%) and 8-OH-dG substrate, the intensity of the color is proportional to the con- increased by about 2-fold (Table 1). In SRS autistic individ- centration of hs-CRP in the sample. A standard curve is pre- uals such differences were not significant. As emerging from pared relating color intensity to the level of the hs-CRP with results shown in Table 1, the marker 8-OH-dG resulted as the hs-CRP undetectable under 0.3 mg/L. more stable in effect either in mild, moderate or in severe ASD, both for CARS and SRS. In order to select the predic- tive variable in the series of parameters considered in Table 1, Assay of 8-hydroxy-2′-deoxyguanosine we fitted different regression models by using an SPSS software using a stepwise regression procedure, in the form of 8-Hydroxy-2′-deoxyguanosine (8-OH-dG) was measured in sequences of F test. Table 2 shows that 8-OH-dG resulted in blood plasma from the patients with ASD and the neurotypical the best independent marker of ASD resulting from the step- controls using the OxiSelect™ Oxidative DNA Damage wise regression analysis on vitamin D deficiency as the de-

ELISA Kit (Cell Biolabs, San Diego, CA, USA). The un- pendent variable. The frequency distribution of 25(OH)D3,in known 8-OH-dG samples or 8-OH-dG standards were first the 28 ASD pediatric subjects matched to the 27 neurotypical added to an 8-OH-dG/bovine serum albumin conjugate children is reported in Fig. 1. preabsorbed microplate. After a brief incubation, an anti-8- It can be easily observed that all control subjects demon- OH-dG mAb was added, followed by an HRP conjugated strate vitamin D levels higher than the known average normal secondary antibody. The 8-OH-dG content in unknown sam- level [56 ng/ml] (Holick et al. 2011), while, in ASD patients, ples was determined by comparison with the predetermined 8- 10/28 demonstrate an insufficient level while 18/28 demon- OH-dG standard curve. Normal 8-OH-dG levels are 4.13 ± strate significantly lower levels than those the normal values 1.01 ng/ml. [56 ng/ml]. 922 Metab Brain Dis (2018) 33:917–931

Table 1 Mean and standard deviation of plasma levels of 25-(OH)D3, hs-CRP,CYP1B1and 8-OH-dG in blood samples drawn from 28 boys with ASD compared to 27 age and sex matched neurotypical children. Data were elaborated following an ANOVA test (Tukey’s post hoc test)

Parameters Group N Mean ± S.D. Percent Change Two-tailed P-value t-test

25(OH)D3 (ng/ml) Control 27 140.43 ± 17.68 100.00 0.001 Autistic patients 28 95.63 ± 26.63 68.10 3.3901 Autism (mild to moderate in CARS) 7 100.61 ± 24.05 71.64 0.046 Autism (severe in CARS) 21 80.68 ± 22.27 57.45 2.0402 Autism (mild to moderate in SRS) 12 95.32 ± 30.79 67.88 0.777 Autism (severe in SRS) 11 98.68 ± 24.52 70.27 0.28400 hs-CRP (mg/l) Control 27 0.59 ± 0.09 100.00 0.001 Autistic patients 28 0.79 ± 0.09 133.90 3.4103 Autism (mild to moderate in CARS) 7 0.78 ± 0.12 132.20 0.745 Autism (severe in CARS) 21 0.79 ± 0.08 133.90 0.3267 Autism (mild to moderate in SRS) 12 0.81 ± 0.10 137.29 0.323 Autism (severe in SRS) 11 0.78 ± 0.08 132.20 0.9967 CYP1B1 (ng/ml) Control 27 0.69 ± 0.07 100.00 0.001 Autistic patients 28 0.48 ± 0.08 69.57 3.4530 Autism (mild to moderate in CARS) 7 0.56 ± 0.09 81.16 0.003 Autism (severe in CARS) 21 0.46 ± 0.07 66.67 2.2302 Autism (mild to moderate in SRS) 12 0.47 ± 0.06 68.12 0.980 Autism (severe in SRS) 11 0.47 ± 0.08 68.12 0.0201 8-OH-dG (ng/ml) Control 27 4.13 ± 1.01 100.00 0.001 Autistic patients 28 8.17 ± 1.04 197.82 3.3921 Autism (mild to moderate in CARS) 7 8.75 ± 0.84 211.86 0.090 Autism (severe in CARS) 21 7.98 ± 1.05 193.22 1.7343 Autism (mild to moderate in SRS) 12 8.00 ± 1.04 193.70 0.393 Autism (severe in SRS) 11 8.37 ± 0.97 202.66 0.9931

Figure 2 shows the frequency distribution of the markers correlated with 25(OH)D3 and CYP1B1, while, on the analyzed in the study. Regarding mild-moderate and severe other hand, 25(OH)D3 was positively correlated with sub-grouping, it can be observed that overlapping was absent CYP1B1 and negatively correlated with hs-CRP and 8- in the case of vitamin D and CYP1B1 but was observed in OH-dG as markers of neuroinflammation and oxidative case of hs-CRP and 8-OH-dG (Fig. 2). If we consider a de- stress respectively (Fig. 4). fined marker for the predictiveness of a binary outcome, such as, e.g., serum concentration of 25(OH)D3, in order to be adapted to calculate the risk of finding ASD in investigated Discussion children in a diagnostic approach, we must calculate a predictiveness curve of the risk model. When we calculated To achieve better outcomes for children with ASD, it is of the predictiveness curves of each marker, any marker ap- critical importance to diagnose the disorder as early as possi- peared to show the ability to predict risk in a quite comparable ble. In the first year of life, many infants show both biological way (Fig. 3). The correlation plots also showed that hs-CRP and behavioral signs that something is abnormal. For other was positively correlated with 8-OH-dG and negatively children with ASD, the autistic symptoms seem to be

Table 2 Stepwise regression analysis in which vitamin D was used as the dependent variable against the three other measured parameters as independent markers, demonstrating that 8-OH-dG was the most contributed marker to 25(OH)D3 deficiency

Predictor variable Beta P value Adjusted R square Modelt

Fvalue Pvalue

8-OH-dG −0.543 0.001 0.288 43.473 0.001 Metab Brain Dis (2018) 33:917–931 923

Fig. 1 Frequency histograms of the plasma calcidiol (25(OH)D3)in each group. It shows the normal distribution of 25(OH)D3 in control and neurotypical controls (panel a) and children with ASD (panel b). ASD patients. It can be easily observed that 21/27 of control subjects Distributions were plotted according to the standard level of reference. demonstrate vitamin D levels higher than the average mean (56 ng/ml), Lognormality curves were fitted on the histograms (black curve) and used while, in ASD patients, 10/28 demonstrate an insufficient level while 18/ to calculate data in a goodness of fitness way (Anderson test and Shapiro- 28 demonstrate significantly lower levels than those of control subjects Wilk test positive, p <0.01).Datafrom28boyswithASDand27ageand (i.e., less than 56 ng/ml) sex matched neurotypical children. A line designates the mean value for clinically absent during infancy, but they regress by age two of the study subjects, even some of the children with (Werner and Dawson 2005). Psychological assessments de- ASD, showed concentrations as high as 40 ng/ml. signed to detect behavioral signs of children as young as However, the normal ranges of vitamin D vary substan- 12 months have shown promise. Examples of this are the tially between different laboratories (Lai et al. 2012). interactive measure Autism Observation Scale for Infants Therefore, it is best to think of them relatively, not (Stone et al. 2004; Bryson et al. 2008), and the First Year categorically. Nevertheless, the patients with ASD in

Inventory, which is a questionnaire for a parent (Baranek the present study had much lower 25(OH)D3 levels et al. 2003; Watson et al. 2007; Turner-Brown et al. 2013). compared to age and gender-matched neurotypical con- However, additional research is needed to assess the clinical trols. This finding is in agreement with previous studies utility of these diagnostic instruments. (Mostafa and Al-Ayadhi 2012;Wangetal.2016).

As already mentioned, 25(OH)D3 levels between 40 The present study found a barely significant association and 60 ng/mL is recommended by the Endocrine between vitamin D levels and autism severity measured by Society (Holick et al. 2011). In the present study, some CARS (P = 0.046). The observation that the levels of 924 Metab Brain Dis (2018) 33:917–931

Fig. 2 a-d Scatter plot points graphs of 25(OH)D3, CYP1B1, hs-CRP, matched neurotypical children in the four measured parameters is noted. and 8-OH-dG. It demonstrates individual data distribution around the Regarding mild-moderate and severe sub-grouping, it can be observed mean value represented by a straight line for the four parameters. The that overlapping was absent in the case of vitamin D and CYP1B1 but absence of overlapping between 28 boys with ASD and 27 age- and sex- was observed in case of hs-CRP and 8-OH-dG

25(OH)D3 are inversely related to autism severity is supported 2014;Chenetal.2016;Saadetal.2016b). Vanlint and Nugent by several studies (Mostafa and Al-Ayadhi 2012;Gongetal. (2006) reported a significant relationship between vitamin D Metab Brain Dis (2018) 33:917–931 925

marker: Cvto P450 (ng/ml) a marker: 25 OH Vit D3 (nmol/L) b 1 1 .8 .8 .6 .6 Risk Risk .4 .4

Prevalence Prevalence .2 .2 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Risk Percentile Risk Percentile

marker: 8-HDG (ng/ml) c marker: CRP (ng/ml) d 1 1 .8 .8 .6 .6 Risk Risk .4 .4

Prevalence Prevalence .2 .2 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Risk Percentile Risk Percentile

Fig. 3 Predictiveness curves of the four measured parameters as an assessment of the performance of 25(OH)D3 CYP1B1, CRP and 8-OH-dG in regards to ASD prevalence in a Saudi population of autistic males deficiency and intellectual disabilities in 337 individuals with outside the blood-brain barrier. According to this mechanism, the cognitive impairments, even after control of the sun exposure. lower level of vitamin D reported in the present study may be a Researchers in Sweden have shown that estrogen makes fe- cause of the significant increase of serotonin seen in the plas- male brains more responsive to neurohormonal growth- ma of patients with ASD (El-Ansary et al. 2011b). Vitamin stimulating effects of calcitriol. This suggests that estrogen en- D may also be related to the remarkable alteration of oxy- hances the beneficial effects of vitamin D on the brain (Fan et al. tocin and vasopressin that has been reported in patients 2006). At the same time, testosterone significantly inhibits with ASD (Xu et al. 2013). Genes that encode the

CYP27B1 (the gene that metabolizes 25(OH)D3 into a steroid oxytocin-neurophysin 1 preproprotein are upregulated by hormone) while stimulating CYP24A1 (the gene responsible for vitamin D. Both oxytocin and vasopressin receptor genes catabolizing vitamin D to be excreted in the bile)(Olmos-Ortiz are to some extent regulated by vitamin D (Olmos-Ortiz et al. 2016). et al. 2016). As estrogen potentiates the effect of vitamin D, and In the present study, elevated levels of hs-CRP in patients testosterone retards it, the differences in the levels of sex with ASD were found compared to neurotypical controls (see steroids during early brain development may show protective also Table 1). This is consistent with at least one previous effects of estrogen on developing female brains due to vitamin study, which demonstrated that hs-CRP might be a biomarker D deficiency, while testosterone cannot protect ASD males as for ASD (Khakzad et al. 2012). Cannell (2008) first hypothesized. The reported insufficiency of vitamin D in ASD could also ThepresentstudyindicatesthatvitaminDmayplayarolein be related to the significant elevations of alpha-N- ASD. About 50% children with ASD have elevated levels of acetylgalactosaminidase (Nagalase) as a marker of secondary serum peripheral serotonin but low central serotonin (Coutinho immune dysregulation in ASD patients compared to controls et al. 2004; Patrick and Ames 2014). Vitamin D in the brain has (Griffin et al. 2003). Much higher Nagalase activity inhibits, been shown to downregulate the transcription of the serotonin- through a deglycosylation reaction, the macrophage activation synthesizing gene tryptophan hydroxylase (TPH) 2 (Patrick and of vitamin D binding protein (DBP), and suppresses the im- Ames 2014). However, vitamin D upregulates TPH1 in tissues mune response. The impact of Nagalase on VDBP 926 Metab Brain Dis (2018) 33:917–931

Fig. 4 These plots demonstrate the correlation between four measured correlated with CRP and 8-OH-dG as markers of neuroinflamation and parameters in 28 boys with ASD and 27 age and sex matched oxidative stress respectively. Statistics obtained from rank groups with a neurotypical children. Hs-CRP was positively correlated with 8-OH-dG Wilcoxon-Mann Whitney test. (Kolmogorov-Smirnov indicated that data and negatively correlated with 25(OH)D3 and CYP1B1. On the other were nonparametric) hand, 25(OH)D3 was positively correlated with CYP1B1 and negatively transportation of vitamin D is not understood. However, the epigenetic silencing of CYP27B1, the gene that encodes for contribution of vitamin D insufficiency to autoimmunity sug- calcitriol, may explain one of the mechanisms in which vita- gests a relationship between elevated Nagalase, GC-MAF min D influences ASD. dysfunction, and vitamin D insufficiency in ASD individuals. Many toxins have been associated with ASD (Lanphear Therefore, Nagalase should be given more attention as a pos- 2015). Toxins are one class of xenobiotics. Other examples sible diagnostic agent. of xenobiotics include drugs, pesticides, cosmetics, flavor- It is well known that cytochrome P450 (CYP) enzymes are ings, fragrances, food additives, industrial chemicals, and en- crucial for the bioactivation of vitamin D in the formation of vironmental pollutants. Humans are exposed to thousands of the secosteroid calcitriol, which requires a 25-hydroxylation xenobiotics in their lifetimes but seemingly catabolize them. followed by a 1α-hydroxylation catalyzed by the liver and Many researchers are on a quest to find xenobiotics that cause kidney cytochrome P450 enzymes (Wikvall 2001). The asso- ASD, without apparently considering the health of the body’s ciation between the levels of CYP1B1 and 25(OH)D3 and systems that excrete xenobiotics. autism severity measured by CARS (P = 0.046 and 0.003 re- Air pollution dramatically reduces vitamin D produced spectively) may show that vitamin D insufficiency and lower from the UVB in sunlight (Agarwal et al. 2002; Windham CYP1B1 activity are both predisposing developmental fac- et al. 2006; Vieira 2015) and has been suggested as a cause tors. To the best of our knowledge, the present study is the of ASD (Hosseinpanah et al. 2010;Baïzetal.2012; Kelishadi first study that reports a significantly lower level of CYP1B1 et al. 2014). If air pollution caused ASD, would not the autism in patients with ASD (Table 1). However, the significance of epidemic in the USA have occurred during the 1950s and 60s this finding is unclear. CYP1B1 may be epigenetically si- when the air pollution was much worse? Perhaps the differ- lenced in ASD. Kouzmenko et al. (2010) hypothesized that ence is that Bin the day^ we were weaned on vitamin D Metab Brain Dis (2018) 33:917–931 927 enriched cow’s milk and played in the sun. Toddlers today are Based on the relationship between vitamin D deficiency protected from sunlight (Gies et al. 2013)andareoften and the increase of DNA damage clinically as evidenced by weaned on unfortified fruit juice (Walker et al. 2006). a significant increase of 8-OH-dG, the remarkable increase in The human body gets rid of xenobiotics, including toxins, autism prevalence could be related to the prevalence of vita- via the CYP3A4 system. CYP3A4 is the most active gene in min D deficiency, which has grown considerably during re- the cytochrome P450 system that codes for detoxifying of cent years, perhaps due to sun avoidance, sunscreen use and xenobiotics. CYP3A4 is very active in the brain (Ferguson an indoor lifestyle (Cannell 2017). The use of sunscreen is and Tyndale 2011) and codes for proteins mainly involved relatively common in pregnant women, to whom strict avoid- in cellular detoxification. CYP3A4 is directly genetically up- ance of sun is often recommended for prevention of chloasma regulated by vitamin D (Thompson et al. 2002;Wangetal. (Lakhdar et al. 2007). 2013). However, ASD children with vitamin D deficiency The suggested relationship between vitamin D deficiency may not be able to upregulate CYP3A4 entirely, and thus their and elevated 8-OH-dG might also be useful as biomarkers for brains are at risk from toxins. Likewise, the body’s master ASD. Peripheral hyperserotonemia in ASD supports the re- antioxidant, glutathione is also directly upregulated by vita- cent work of Alfawaz et al. (2014) who demonstrated signif- min D (Rose et al. 2012; Jain and Micinski 2013). Glutathione icant protective and restorative effects of vitamin D on low has been shown to be low in autistic cerebellums (Jain and serotonin, interferon-gamma (IFγ) and glutathione-s- Micinski 2013). The administration of vitamin D significantly transferase (GST) induced by the administration of propionic increases serum glutathione levels (Zhang et al. 2016). (PPA) acid to produce a rodent model of autism. Therefore, impaired defense may contribute to ASD, and not While the early ASD diagnosis is important for the institu- just the toxins. tion of behavioral therapy, a recent randomized controlled trial The neuroinflammation in ASD (El-Ansary and Al-Ayadhi (RCT), published in the Journal of Child Psychiatry and 2012; Bjørklund et al. 2016) may be caused by chronic over- Psychology, found high dose vitamin D (300 IU/KG/day) production of pro-inflammatory cytokines and oxidative stress had a significant treatment effect on the core symptoms of occurring in genetically susceptible vitamin D deficient chil- ASD (Saad et al. 2016b). The authors also reported that youn- dren (Mazahery et al. 2016). Very severe inflammation, such ger children and those children whose final 25(OH)D3 as sepsis, significantly depresses the activity of several en- exceeded >40 ng/ml responded best. The finding of this zymes including the cytochrome CYP450 family responsible RCT is supported by positive open-label trials (Cannell for the catabolism of xenobiotics (Jacob et al. 2009). Also, at 2017). If replicated, it would make the early diagnosis of least one other study found hs-CRP is elevated in ASD and ASD critical, because effective treatment with vitamin D may be used by clinicians to help diagnose ASD (El-Ansary may be possible. et al. 2011b). The reported decrease in detoxification capacity in Saudi ASD patients reported by Al-Yafee et al. (2011)may contribute to the etiology of ASD and is consistent with the Limitations finding of elevated inflammatory markers in the present study and the severity of autism as measured by the CARS and SRS. The findings of the present study are hindered by a relatively DNA damage is an underlying cause of much cellular dys- small sample size, and will, therefore, need to be replicated by function. If the DNA goes unrepaired, it can drive mutagene- other groups before clinical utilization. Also, all the subjects in sis that disrupts normal gene expression and produces proteins this study were males, so the findings cannot be assumed to be with abnormal functions. 8-OH-dG is the most abundant prod- present in females with ASD. The 25(OH)D3 levels of both uct of cellular DNA oxidation (Prabhulkar and Li 2010). the subjects and controls in the present study were much Table 1 demonstrates the significant increase of 8-OH-dG in higher than in other studies of children with ASD. Saudi ASD patients compared to neurotypical control subjects. Arabia is relatively close to the equator and has a sunny cli-

Also, multiple regression analysis using 25(OH)D3 as mate, which may explain some of the findings of the present the dependent variable (Table 2) demonstrates a correla- study. Also, operator error is a possibility. In addition, as cited tion between vitamin D deficiency and the significant above, 25(OH)D3 levels vary from lab to lab and should be increase of 8-OH-dG. This may support the recent work thought of relatively rather than absolutely. of Gordon-Thomson et al. (2012) in which they reported that DNA strand breaks in UV-irradiated human keratinocytes where 8-oxoguanine DNA glycosylase di- Conclusions gests increased more than 2-fold but was decreased by calcitriol treatment. An analog of calcitriol significantly The current problem with biomarkers is that numerous small reduced 8-OH-dG levels in rats who had a cyclosporine- studies have found multiple, but different, biomarkers for induced renal injury (Piao et al. 2012). ASD. Researchers studying biomarkers for ASD need to meet 928 Metab Brain Dis (2018) 33:917–931 and collaborate more as it appears that a meta-analysis cannot Anderson DK, Liang JW, Lord C (2014) Predicting young adult outcome be done now because such a wide variety of biomarkers have among more and less cognitively able individuals with autism spectrum disorders. J Child Psychol Psychiatry 55:485–494 been studied, and so a useful meta-analysis cannot be done. APA-American Psychiatric Association (2000) Diagnostic and statistical The present study demonstrates lower vitamin D levels in manual of mental disorders: DSM-IV-TR. American Psychiatric Saudi children with ASD, which agrees with a recent meta- Association, Washington, DC analysis as mentioned above. We found a significant inverse Baïz N, Dargent-Molina P, Wark JD, Souberbielle JC, Slama R, Annesi- Maesano I, EDEN Mother-Child Cohort Study Group (2012) association between hypovitaminosis D and impaired cogni- Gestational exposure to urban air pollution related to a decrease in tive development as measured on the CARS. This is compat- cord blood vitamin D levels. J Clin Endocrinol Metab 97:4087– ible with the hypothesis that vitamin D plays a role in the 4095 etiology of ASD. We replicated the reports of others in finding Baranek GT, Watson LR, Crais E, Reznick S (2003) First-Year Inventory (FYI) 2.0. University of North Carolina, Chapel Hill that 25(OH)D3 and hs-CRP are abnormal in ASD. Two pre- Bikle DD, Siiteri PK, Ryzen E et al (1985) Serum protein binding of 1,25- vious studies (see above) found that 8-OH-dG was not elevat- dihydroxyvitamin D: a reevaluation by direct measurement of free ed in ASD. However, it was elevated in the sample of the metabolite levels. J Clin Endocrinol Metab 61:969–975 present study. As far as we know, we are the first to report Bjørklund G (2016) Vitamin D deficiency: a global health problem. Peertechz J Environ Sci Toxicol 1:23–24 https://www.peertechz. that CYP1B1 is reduced in ASD. com/Environmental-Science-Toxicology/pdf/PJEST-1-104.pdf. Further studies should be conducted with a larger number Accessed 21 April 2017 of participants to confirm these preliminary findings and to Bjørklund G, Saad K, Chirumbolo S, Kern JK, Geier DA, Geier MR, advance towards the goal of a set of biomarkers that will lead Urbina MA (2016) Immune dysfunction and neuroinflammation in autism spectrum disorder. Acta Neurobiol Exp (Wars) 76:257–268 to earlier ASD diagnosis and intervention. Bryson SE, Zwaigenbaum L, McDermott C, Rombough V, Brian J (2008) The autism observation scale for infants: scale development Acknowledgements This research project was supported by a grant from and reliability data. J Autism Dev Disord 38:731–738 the Research Center of the Center for Female Scientific and Medical Calderón-Garcidueñas L, Wen-Wang L, Zhang YJ, Rodriguez-Alcaraz A, Colleges at King Saud University. Osnaya N, Villarreal-Calderón A, Santella RM (1999) 8-hydroxy-2'- deoxyguanosine, a major mutagenic oxidative DNA lesion, and Compliance with ethical standards DNA strand breaks in nasal respiratory epithelium of children exposed to urban pollution. Environ Health Perspect 107:469–474 Cannell JJ (2008) Autism and vitamin D. Med Hypotheses 70:750–759 Conflict of interest The authors declare no potential conflicts of interest Cannell JJ (2013) Autism, will vitamin D treat core symptoms? Med with respect to the authorship, and/or publication of this article. Hypotheses 81:195–198 Cannell JJ (2017) Vitamin D and autism, what’s new? Rev Endocr Metab Ethical approval All procedures performed were in accordance with the Disord. https://doi.org/10.1007/s11154-017-9409-0 ethical standards of the institutional and/or national research committee, Carpenter TO, Zhang JH, Parra E, Ellis BK, Simpson C, Lee WM, Balko and with the 1964 Helsinki declaration and its later amendments or com- J, Fu L, Wong BY, Cole DE (2013) Vitamin D binding protein is a parable ethical standards. key determinant of 25-hydroxyvitamin D levels in infants and toddlers. J Bone Miner Res 28:213–221 Chakrabarti B, Dudbridge F, Kent L, Wheelwright S, Hill-Cawthorne G, Allison C, Banerjee-Basu S, Baron-Cohen S (2009) Genes Related to Sex Steroids, Neural Growth, and Social–Emotional Behavior are References Associated with Autistic Traits, Empathy, and Asperger Syndrome. 9 International Society for Autism Research, Wiley Periodicals, Inc. – ADDM Network-Autism and Developmental Disabilities Monitoring p1 21 Network Surveillance Year 2010 Principal Investigators (2014) Chauhan A, Chauhan V, Brown WT, Cohen I (2004) Oxidative stress in Prevalence of autism spectrum disorder among children aged 8 years autism: increased lipid peroxidation and reduced serum levels of - autism and developmental disabilities monitoring network, 11 ceruloplasmin and transferring the antioxidant proteins. Life Sci – sites, United States, 2010. MMWR Surveill Summ 63(2):1–21 75:2539 2549 Chen J, Xin K, Wei J, Zhang K, Xiao H (2016) Lower maternal serum Agarwal KS, Mughal MZ, Upadhyay P, Berry JL, Mawer EB, Puliyel JM 25(OH) D in first trimester associated with higher autism risk in (2002) The impact of atmospheric pollution on vitamin D status of Chinese offspring. J Psychosom Res 89:98–101 infants and toddlers in Delhi, India. Arch Dis Child 87:111–113 Cherniack EP, Troen BR (2011) Vitamin D, cognitive function, and men- Alfawaz HA, Shafi Bhat R, Al-Ayadhi L, El-Ansary AK (2014) tal health. In: Benton D (ed) Lifetime nutritional influences on cog- Protective and restorative potency of Vitamin D on persistent bio- nition, behaviour and psychiatric illness. Woodhead Publishing, chemical autistic features induced in propionic acid-intoxicated rat Sawston, pp 420– 438. https://doi.org/10.1533/9780857092922.3. pups. BMC Complement Altern Med 14:416. https://doi.org/10. 420 1186/1472-6882-14-416 Chirumbolo S, Bjørklund G, Sboarina A, Vella A (2017) The role of Ali A, Cui X, Eyles D (2016) Developmental vitamin D deficiency and vitamin D in the immune system as a pro-survival molecule. Clin autism: putative pathogenic mechanisms. J Steroid Biochem Mol Ther. https://doi.org/10.1016/j.clinthera.2017.03.021 Biol. https://doi.org/10.1016/j.jsbmb.2016.12.018 Christensen DL, Baio J, Van Naarden BK, Bilder D, Charles J, Al-Yafee YA, Al-Ayadhi LY, Haq SH, El-Ansary AK (2011) Novel met- Constantino JN, Daniels J, Durkin MS, Fitzgerald RT, Kurzius- abolic biomarkers related to sulfur-dependent detoxification path- Spencer M, Lee LC, Pettygrove S, Robinson C, Schulz E, Wells ways in autistic patients of Saudi Arabia. BMC Neurol 11:139. C, Wingate MS, Zahorodny W, Yeargin-Allsopp M, Centers for https://doi.org/10.1186/1471-2377-11-139 Disease Control and Prevention (CDC) (2016) Prevalence and Metab Brain Dis (2018) 33:917–931 929

characteristics of autism spectrum disorder among children aged 8 Ferguson CS, Tyndale RF (2011) Cytochromes P450 in the brain: years–Autism and Developmental Disabilities Monitoring Network, Emerging evidence for biological significance. Trends Pharmacol 11 sites, United States, 2012. MMWR Surveill Summ 65(3):1–23. Sci 32:708–714 https://doi.org/10.15585/mmwr.ss6503a1 Fernell E, Bejerot S, Westerlund J, Miniscalco C, Simila H, Eyles D, Chun SK, Shin SA, Kim MY, Joung H, Chung J (2016) Effects of ma- Gillberg C, Humble MB (2015) Autism spectrum disorder and ternal genetic polymorphisms in vitamin D-binding protein and se- low vitamin D at birth: a sibling control study. Mol Autism 6:3. rum 25-hydroxyvitamin D concentration on infant birth weight. https://doi.org/10.1186/2040-2392-6-3 Nutrition 35:36–42. https://doi.org/10.1016/j.nut.2016.10.006 Fukuda M, Yamaguchi H, Yamamoto H, Aminaka M, Muramaka H, Constantino JN, Davis SA, Todd RD, Schindler MK, Gross MM, Brophy Kamiyama N et al (2008) The evaluation of oxidative stress damage SL, Metzger LM, Shoushtari CS, Splinter R, Reich W (2003) in children with brain damage using 8 OH deoxyguanosine levels. Validation of a brief quantitative measure of autistic traits: compar- Brain Dev 30:131–136 ison of the social responsiveness scale with the autism diagnostic Garcion E, Thanh XD, Bled F, Teissier E, Dehouck MP, Rigault F, interview-revised. J Autism Dev Disord 33:427–433 Brachet P, Girault A, Torpier G, Darcy F (1996) 1,25- Coutinho AM, Oliveira G, Morgadinho T, Fesel C, Macedo TR, Bento C, Dihydroxyvitamin D3 regulates gamma 1 transpeptidase activity Marques C, Ataíde A, Miguel T, Borges L, Vicente AM (2004) in rat brain. Neurosci Lett 216:183–186 Variants of the serotonin transporter gene (SLC6A4) significantly Garcion E, Wion-Barbot N, Montero-Menei CN, Berger F, Wion D contribute to hyperserotonemia in autism. Mol Psychiatry 9:264– (2002) New clues about vitamin D functions in the nervous system. 271 Trends Endocrinol Metab 13:100–105 Currenti SA (2010) Understanding and determining the etiology of au- Ghanizadeh A, Akhondzadeh S, Hormozi M, Makarem A, Abotorabi- – tism. Cell Mol Neurobiol 30:161 171 Zarchi M, Firoozabadi A (2012) Glutathione-related factors and Dawodu A, Wagner CL (2012) Prevention of vitamin D deficiency in oxidative stress in autism, a review. Curr Med Chem 19:4000–4005 mothers and infants worldwide - a paradigm shift. Paediatr Int Child Gies P, Makin J, Dobbinson S, Javorniczky J, Henderson S, Guilfoyle R, – Health 32:3 13 Lock J (2013) Shade provision for toddlers at swimming pools in Dawodu A, Zalla L, Woo JG, Herbers PM, Davidson BS, Heubi JE, Melbourne. Photochem Photobiol 89:968–973 Morrow AL (2014) Heightened attention to supplementation is Gong ZL, Luo CM, Wang L, Shen L, Wei F, Tong RJ, Liu Y (2014) needed to improve the vitamin D status of breastfeeding mothers Serum 25-hydroxyvitamin D levels in chinese children with autism and infants when sunshine exposure is restricted. Matern Child Nutr spectrum disorders. Neuroreport 25:23–27 10:383–397 Gonzalez-Suarez I, Redwood AB, Grotsky DA, Neumann MA, Cheng Delanghe JR, Speeckaert R, Speeckaert MM (2015) Behind the scenes of EH, Stewart CL, Dusso A, Gonzalo S (2011) A new pathway that vitamin D binding protein: more than vitamin D binding. Best Pract regulates 53BP1 stability implicates cathepsin L and vitamin D in Res Clin Endocrinol Metab 29:773–786 DNA repair. EMBO J 30:3383–3396. https://doi.org/10.1038/ Dey A (2013) Cytochrome P450 2E1: its clinical aspects and a brief emboj.2011.225 perspective on the current research scenario. Subcell Biochem 67: Gordon-Thomson C, Gupta R, Tongkao-on W, Ryan A, Halliday GM, 1–104 Mason RS (2012) 1α,25 Dihydroxyvitamin D3 enhances cellular Du L, Shan L, Wang B, Feng JY, Xu ZD, Jia FY (2015) Serum levels of defenses against UV-induced oxidative and other forms of DNA 25-hydroxyvitamin D in children with autism spectrum disorders. damage in skin. Photochem Photobiol Sci 11:1837–1847 Zhongguo Dang Dai Er Ke Za Zhi 17:68–71 Graziano S, Johnston R, Deng O, Zhang J, Gonzalo S (2016) Vitamin D/ El-Ansary A (2016) Data of multiple regressions analysis between select- vitamin D receptor axis regulates DNA repair during oncogene- ed biomarkers related to glutamate excitotoxicity and oxidative induced senescence. Oncogene 35:5362–5376 stress in Saudi autistic patients. Data Brief 7:111–116 El-Ansary A, Al-Ayadhi L (2012) Neuroinflammation in autism spectrum Griffin MD, Xing N, Kumar R (2003) Vitamin D and its analogs as disorders. J Neuroinflammation 9:265. https://doi.org/10.1186/ regulators of immune activation and antigen presentation. Annu – 1742-2094-9-265 Rev Nutr 23:117 145 El-Ansary AK, Bacha AG, Al-Ayahdi LY (2011a) Plasma fatty acids as Gross ML, Tenenbein M, Sellers EA (2013) Severe vitamin D deficiency diagnostic markers in autistic patients from Saudi Arabia. Lipids in 6 Canadian first nation formula-fed infants. Int J Circumpolar Health Dis 10:62. https://doi.org/10.1186/1476-511X-10-62 Health 72:20244. https://doi.org/10.3402/ijch.v72i0.20244 El-Ansary A, Ben Bacha AG, Al-Ayadhi LY (2011b) Proinflammatory Helbock HJ, Beckman KB, Ames BN (1999) 8 hydroxy deoxyguanosine and proapoptotic markers in relation to mono and di-cations in plas- and 8hydroxyguanine as biomarkers of oxidative DNA damage. – ma of autistic patients from Saudi Arabia. J Neuroinflammation 8: Methods Enzymol 300:156 166 142. https://doi.org/10.1186/1742-2094-8-142 Hodgson NW, Waly MI, Al-Farsi YM, Al-Sharbati MM, Al-Farsi O, Ali El-Ansary A, Hassan WM, Qasem H, Das UN (2016) Identification of A, Ouhtit A, Zang T, Zhou ZS, Deth RC (2014) Decreased glutathi- biomarkers of impaired sensory profiles among autistic patients. one and elevated hair mercury levels are associated with nutritional PLoS One 11:e0164153. https://doi.org/10.1371/journal.pone. deficiency-based autism in Oman. Exp Biol Med (Maywood) 239: 0164153 697–706 Emond P, Mavel S, Aïdoud N et al (2013) GC-MS-based urine metabolic Holick MF (2006) Resurrection of vitamin D deficiency and rickets. J profiling of autism spectrum disorders. Anal Bioanal Chem 405: Clin Invest 116:2062–2072 5291–5300 Holick MF (2007) Vitamin D deficiency. N Engl J Med 357:266–281 Fan X, Warner M, Gustafsson JA (2006) Estrogen receptor beta expres- Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, sion in the embryonic brain regulates development of calretinin- Heaney RP, Murad MH, Weaver CM, Endocrine Society (2011) immunoreactive GABAergic interneurons. Proc Natl Acad Sci U S Evaluation, treatment, and prevention of vitamin D deficiency: an A. 103:19338–19343 Endocrine Society clinical practice guideline. J Clin Endocrinol Fedirko V, Bostick RM, Long Q, Flanders WD, Fedirko V, Bostick RM, Metab 96:1911–1930 Long Q, Flanders WD, McCullough ML, Sidelnikov E et al (2010) Hosseinpanah F, Pour SH, Heibatollahi M, Moghbel N, Asefzade S, Effects of supplemental vitamin D and calcium on oxidative DNA Azizi F (2010) The effects of air pollution on vitamin D status in damage marker in normal colorectal mucosa: a randomized clinical healthy women: a cross sectional study. BMC Public Health 10:519. trial. Cancer Epidemiol Biomarkers Prev 19:280–291 https://doi.org/10.1186/1471-2458-10-519 930 Metab Brain Dis (2018) 33:917–931

Huang Y, Sullivan Pepe M, Feng Z (2007) Evaluating the predictiveness Mahmoud AA, Ali AH (2014) Vitamin D receptor gene polymorphism of a continuous marker. Biometrics. 63(4):1181–1188 and 25 hydroxy vitamin D levels in Egyptian patients with pulmo- Jacob A, Zhou M, Wu R, Wang P (2009) The role of hepatic cytochrome nary tuberculosis. Egypt J Chest Dis Tuberc 63:651–655 P-450 in sepsis. Int J Clin Exp Med 2:203–211 Marusteri M, Bacarea V (2010) Comparing groups for statistical differ- Jain SK, Micinski D (2013) Vitamin D upregulates glutamate cysteine ences: how to choose the right statistical test? Biochemia Medica 20: ligase and glutathione reductase, and GSH formation, and decreases 15–32 ROS and MCP-1 and IL-8 secretion in high-glucose exposed U937 Mazahery H, Camargo CA, Conlon C, Beck KL, Kruger MC, von Hurst monocytes. Biochem Biophys Res Commun 437:7–11 PR (2016) Vitamin D and autism spectrum disorder: a literature James SJ (2013) Autism and folate-dependent one-carbon metabolism: review. Nutrients 8:236. https://doi.org/10.3390/nu8040236 serendipity and critical branch-point decisions in science. Glob Adv Mellenthin L, Wallaschofski H, Grotevendt A, Völzke H, Nauck M, Health Med 2:48–51 Hannemann A (2014) Association between serum vitamin D con- James SJ, Cutler P, Melnyk S, Jernigan S, Janak L, Gaylor DW, centrations and inflammatory markers in the general adult popula- Neubrander JA (2004) Metabolic biomarkers of increased oxidative tion. Metabolism 63:1056–1062 stress and impaired methylation capacity in children with autism. Mick KA (2005) Diagnosing autism: Comparison of the Childhood Am J Clin Nutr 80:1611–1617 Autism Rating Scale (CARS) and the Autism Diagnostic Jiménez-Garza O, Baccarelli AA, Byun HM, Márquez-Gamiño S, Observation Schedule (ADOS). Dissertation, Wichita State Barrón-Vivanco BS, Albores A (2015) CYP2E1 epigenetic regula- University tion in chronic, low-level toluene exposure: Relationship with oxi- Ming X, Stein TP, Brimacombe M, Johnson WG, Lambert GH, Wagner dative stress and smoking habit. Toxicol Appl Pharmacol 286:207– GC (2005) Increased excretion of a lipid peroxidation biomarker in 215 autism. Prostaglandins Leukot Essent Fatty Acids 73:379–384 Jones KS, Assar S, Vanderschueren D, Bouillon R, Prentice A, Ming X, Stein TP, Barnes Vet al (2012) Metabolic perturbance in autism Schoenmakers I (2015) Predictors of 25(OH)D3 half-life and plas- spectrum disorders: a metabolomics study. J Proteome Res 11: ma 25(OH)D3 concentration in Gambia and the UK. Osteoporos Int 5856–5862 26:1137–1146 Mostafa GA, Al-Ayadhi LY (2012) Reduced serum concentrations of 25- Kelishadi R, Moeini R, Poursafa P, Farajian S, Yousefy H, Okhovat- hydroxy vitamin D in children with autism: relation to autoimmuni- Souraki AA (2014) Independent association between air pollutants ty. J Neuroinflammation 9:201. https://doi.org/10.1186/1742-2094- and vitamin D deficiency in young children in Isfahan, Iran. Paediatr 9-201 Int Child Health 34:50–55 Nebert DW, Russell DW (2002) Clinical importance of the cytochromes Khakzad MR, Javanbakht M, Shayegan MR, Kianoush S, Omid F, Hojati P450. Lancet 360:1155–1162 M, Meshkat M (2012) The complementary role of high sensitivity Olmos-Ortiz A, García-Quiroz J, López-Marure R, González-Curiel I, C-reactive protein in the diagnosis and severity assessment of au- Rivas-Santiago B, Olivares A, Avila E, Barrera D, Halhali A, tism. Res Autism Spectr Disord 6:1032–1037 Caldiño F, Larrea F, Díaz L (2016) Evidence of sexual dimorphism Kinney DK, Barch DH, Chayka B, Napoleon S, Munir KM (2010) in placental vitamin D metabolism: Testosterone inhibits calcitriol- Environmental risk factors for autism: do they help cause de novo dependent cathelicidin expression. J Steroid Biochem Mol Biol 163: genetic mutations that contribute to the disorder? Med Hypotheses 173–182 74:102–106 Patrick RP, Ames BN (2014) Vitamin D hormone regulates serotonin Kouzmenko A, Ohtake F, Fujiki R, Kato S (2010) Hormonal gene regu- synthesis. Part 1: relevance for autism. FASEB J 28:2398–2413 lation through DNA methylation and demethylation. Epigenomics Pelphrey K (2017) Charting a course for autism biomarkers. Biol 2:765–774 Psychiatry 82:155–156 Kuwabara H, Yamasue H, Koike S, Inoue H, Kawakubo Y, Kuroda M, Pepe MS, Feng Z, Huang Y, Longton G, Prentice R, Thompson IM, Takano Y,Iwashiro N, Natsubori T, Aoki Y,Kano Y,Kasai K (2013) Zheng Y (2008) Integrating the predictiveness of a marker with its Altered metabolites in the plasma of autism spectrum disorder: a performance as a classifier. Am J Epidemiol. 167:362–368 capillary electrophoresis time-of-flight mass spectroscopy study. Piao SG, Song JC, Lim SW, Chung BH, Choi BS, Yang CW (2012) PLoS ONE 8:e73814. https://doi.org/10.1371/journal.pone. Protective effect of paricalcitol on cyclosporine-induced renal injury 0073814 in rats. Transplant Proc 44:642–645 Lai JK, Lucas RM, Banks E, Ponsonby AL (2012) Ausimmune investi- Prabhulkar S, Li CZ (2010) Assessment of oxidative DNA damage and gator group. Variability in vitamin D assays impairs clinical assess- repair at single cellular level via real-time monitoring of 8-OHdG ment of vitamin D status. Intern Med J 42:43–50 biomarker. Biosens Bioelectron 26:1743–1749 Lakhdar H, Zouhair K, Khadir K, Essari A, Richard A, Seité S, Rougier A Rose S, Melnyk S, Pavliv O, Bai S, Nick TG, Frye RE, James SJ (2012) (2007) Evaluation of the effectiveness of a broad-spectrum sun- Evidence of oxidative damage and inflammation associated with screen in the prevention of chloasma in pregnant women. J Eur low glutathione redox status in the autism brain. Transl Psychiatry Acad Dermatol Venereol 21:738–742 10:e134. https://doi.org/10.1038/tp.2012.61 Lanphear BP (2015) The impact of toxins on the developing brain. Annu Saad K, Abdel-Rahman AA, Elserogy YM, Al-Atram AA, Cannell JJ, Rev Public Health 36:211–230 Bjørklund G, Abdel-Reheim MK, Othman HA, El-Houfey AA, Abd Lin XF, Wang ZY, Feng WJ (2015) Influence of drug-metabolizing en- El-Aziz NH, Abd El-Baseer KA, Ahmed AE, Ali AM (2016a) zymes by autism treatment drug haloperidol. Latin American Vitamin D status in autism spectrum disorders and the efficacy of Journal of Pharmacy 34:1693–1696 vitamin D supplementation in autistic children. Nutr Neurosci 19: Løken-Amsrud KI, Holmøy T, Bakke SJ, Beiske AG, Bjerve KS et al 346–351 (2012) Vitamin D and disease activity in multiple sclerosis before Saad K, Abdel-Rahman AA, Elserogy YM, Al-Atram AA, El-Houfey and during interferon-β treatment. Neurology 79:267–273 AA, Othman HA, Bjørklund G, Jia F, Urbina MA, Abo-Elela MG, Magnusson C, Lundberg M, Lee BK, Rai D, Karlsson H, Gardner R, Ahmad FA, Abd El-Baseer KA, Ahmed AE, Abdel-Salam AM Kosidou K, Arver S, Dalman C (2016) Maternal vitamin D deficien- (2016b) Randomized controlled trial of vitamin D supplementation cy and the risk of autism spectrum disorders: population-based in children with autism spectrum disorder. J Child Psychol study. BJPsych Open 2:170–172 Psychiatry. https://doi.org/10.1111/jcpp.12652 Metab Brain Dis (2018) 33:917–931 931

Sajdel-Sulkowska E, Lipinski B, Windom H, Audhya J, Mc Ginis W Biochem Mol Biol 136:54–58. https://doi.org/10.1016/j.jsbmb. (2008) Oxidative stress in autism: cerebelar 3 nitroryrosine levels. 2012.09.012 Am J Biochem Biotechnol 4:73–84 Wang T, Shan L, Du L, Feng J, Xu Z, Staal WG, Jia F (2016) Serum Sajdel-Sulkowska EM, Xu M, Koibuchi N (2009) Increase in cerebellar concentration of 25-hydroxyvitamin D in autism spectrum disorder: 3-neurotrophin and oxidative stress markers in autism. Cerebellum a systematic review and meta-analysis. Eur Child Adolesc 8:366–372 Psychiatry 25:341–350 Salhia HO, Al-Nasser LA, Taher LS, Al-Khathaami AM, El-Metwally Watson LR, Baranek GT, Crais ER, Steven RJ, Dykstra J, Perryman T AA (2014) Systemic review of the epidemiology of autism in Arab (2007) The first year inventory: retrospective parent responses to a Gulf countries. Neurosciences (Riyadh) 19:291–296 questionnaire designed to identify one-year-olds at risk for autism. J Saltytė Benth J, Myhr KM, Løken-Amsrud KI, Beiske AG, Bjerve KS, Autism Dev Disord 37:49–61 Hovdal H, Midgard R, Holmøy T (2012) Modelling and prediction Wegienka G, Kaur H, Sangha R, Cassidy-Bushrow AE (2016) Maternal- of 25-hydroxyvitamin D levels in Norwegian relapsing-remitting cord blood Vitamin D correlations vary by maternal levels. J – multiple sclerosis patients. Neuroepidemiology 39:84 93 Pregnancy 2016:7474192. https://doi.org/10.1155/2016/7474192 Schuster I (2011) Cytochromes P450 are essential players in the vitamin Werner E, Dawson G (2005) Validation of the phenomenon of autistic – D signaling system. Biochim Biophys Acta 1814:186 199 regression using home videotapes. Arch Gen Psychiatry 62:889– Singh VK (2005) Elevation of serum c-reactive protein and s100 proteins 895 for systemic inflammation in autistic children. J Spec Educ Rehab 6: Wikvall K (2001) Cytochrome P450 enzymes in the bioactivation of 117–135 vitamin D to its hormonal form (review). Int J Mol Med 7:201–209 Siniscalco D, Bradstreet JJ, Cirillo A, Antonucci N (2014) The in vitro GcMAF effects on endocannabinoid system transcriptionomics, re- Windham GC, Zhang L, Gunier R, Croen LA, Grether JK (2006) Autism ceptor formation, and cell activity of autism-derived macrophages. J spectrum disorders in relation to distribution of hazardous air pollutants in the San Francisco Bay area. Environ Health Perspect 114: Neuroinflammation 11:78. https://doi.org/10.1186/1742-2094-11- – 78 1438 1444 Stone WL, Coonrod EE, Turner LM, Pozdol SL (2004) Psychometric Xu XJ, Shou XJ, Li J, Jia MX, Zhang JS, Guo Y,Wei QY,Zhang XT, Han properties of the STAT for early autism screening. J Autism Dev SP, Zhang R, Han JS (2013) Mothers of autistic children: lower Disord 34:691–701 plasma levels of oxytocin and Arg-vasopressin and a higher level Thompson PD, Jurutka PW, Whitfield GK, Myskowski SM, Eichhorst of testosterone. PLoS One 8:e74849. https://doi.org/10.1371/ KR, Dominguez CE, Haussler CA, Haussler MR (2002) Liganded journal.pone.0074849 VDR induces CYP3A4 in small intestinal and colon cancer cells via Yang CJ, Liu CL, Sang B, Zhu XM, Du YJ (2015) The combined role of DR3 and ER6 vitamin D responsive elements. Biochem Biophys serotonin and interleukin-6 as biomarker for autism. Neuroscience Res Commun 299:730–738 284:290–296 Turner-Brown LM, Baranek GT, Reznick JS, Watson LR, Crais ER Yui K, Tanuma N, Yamada H, Kawasaki Y (2016) Reduced endogenous (2013) The first year inventory: a longitudinal follow-up of 12- urinary total antioxidant power and its relation of plasma antioxidant montholds to 3 years of age. Autism 17:527–540 activity of superoxide dismutase in individuals with autism spectrum Vanlint S, Nugent M (2006) Vitamin D and fractures in people with disorder. Int J Dev Neurosci. https://doi.org/10.1016/j.ijdevneu. intellectual disability. J Intellect Disabil Res 50(Pt 10):761–767. 2016.08.003 https://doi.org/10.1111/j.1365-2788.2006.00841.x Yusuf A, Elsabbagh M (2015) At the cross-roads of participatory research Vickers AJ, Elkin EB (2006) Decision curve analysis: a novel method for and biomarker discovery in autism: the need for empirical data. evaluating prediction models. Med Decis Making. 26(6):565–574 BMC Med Ethics 16:88. https://doi.org/10.1186/s12910-015- Vieira SE (2015) The health burden of pollution: the impact of prenatal 0082-0 exposure to air pollutants. Int J Chron Obstruct Pulmon Dis 10: Zablotsky B, Black LI, Maenner MJ, Schieve LA, Blumberg SJ (2015) 1111–1121 Estimated prevalence of autism and other developmental disabilities Vinkhuyzen AA, Eyles DW, Burne TH, Blanken LM, Kruithof CJ, following questionnaire changes in the 2014 National Health Verhulst F, Jaddoe VW, Tiemeier H, McGrath JJ (2016) Interview Survey. Natl Health Stat Report 87:1–20 Gestational vitamin D deficiency and autism-related traits: the Zaky EA, Fouda EM, Algohary EA, Al-Shony ES (2015) Prevalence of Generation R Study. Mol Psychiatry. https://doi.org/10.1038/mp. autism spectrum disorders in vitamin D deficient or insufficient 2016.213 rickets. Int J Sci Res (Raipur) 4:1365–1373 ’ Walker RB, Conn JA, Davies MJ, Moore VM (2006) Mothers views on Zhang Q, Cheng Y, He M, Li T, Ma Z, Cheng H (2016) Effect of various feeding infants around the time of weaning. Public Health Nutr 9: doses of vitamin D supplementation on pregnant women with ges- – 707 713 tational diabetes mellitus: A randomized controlled trial. Exp Ther Wang Z, Schuetz EG, Xu Y, Thummel KE (2013) Interplay between Med 12:1889–1895 vitamin D and the drug metabolizing enzyme CYP3A4. J Steroid